B5 · Publication Volume 10

Closure Temperature and Resetting

open and closed systems, inheritance, loss and mixed ages

Learning goals

After this lesson, you should be able to explain closure as an emergent kinetic boundary, identify the parameters in a diffusion-based closure model, calculate a diffusion length, and distinguish volume diffusion from dissolution-reprecipitation, deformation-assisted exchange and fluid-mediated resetting. You should be able to relate age zoning to grain geometry and thermal history.

The key correction to a common shortcut is that a mineral does not possess one immutable closure temperature. Closure depends on isotope system, diffusion parameters, grain size and geometry, cooling rate, composition, defects, damage and boundary conditions. A quoted value is a model result for stated conditions, not a mineral property like atomic number.

Closure as a kinetic boundary

Closure and resetting model linking cooling rate, grain size, diffusion, fluid reaction and preserved age domains
Closure and resetting model linking cooling rate, grain size, diffusion, fluid reaction and preserved age domains

At high temperature, parent or daughter species may exchange rapidly enough that a mineral tracks its surroundings. As temperature falls, diffusion slows and the system progressively retains radiogenic products. Closure temperature approximates the temperature associated with the effective transition for a specified cooling history and diffusion model. Closure is distributed through time and space rather than an instantaneous physical switch.

Diffusion commonly follows an Arrhenius relation,

$D=D_0\exp\left(-\frac{E}{RT}\right),$

where D_0 is a pre-exponential factor, E activation energy, R the gas constant and T absolute temperature. Small changes in temperature or uncertain E can change D strongly. Composition, pressure, water, radiation damage and crystallographic direction may alter parameters.

A diffusion-based closure expression relates activation energy, diffusion geometry, effective grain dimension, cooling rate and closure temperature through a logarithmic term. The result must be solved consistently with the thermal path. Transplanting a tabulated temperature without its radius and cooling-rate assumptions discards the model.

Grain geometry and diffusion domains

The relevant diffusion dimension is not automatically the hand-picked grain diameter. It may be the radius of a coherent crystal domain, distance to a fast pathway, thickness of a rim, or spacing between fractures. Anisotropic minerals can have different diffusivity by direction. Irregular grains and zoned diffusion coefficients require numerical or bounded models.

A characteristic diffusion length is

$\ell\approx\sqrt{Dt}.$

This scaling helps compare exchange distance with domain size, but D changes strongly during a thermal path. A time-integrated model uses \int D(t)dt. If \ell is small relative to radius, a core may retain older information while a rim exchanges. If comparable, partial resetting and smooth or sharp age gradients may develop depending on boundary conditions and reaction.

Multi-domain behaviour can occur within one mineral separate. Grain-size spectra, subgrains, defects and alteration mean that one bulk date can mix different retentivities. Step-heating or spatial analysis may reveal domains, but inverse models are non-unique unless petrography and thermal constraints are included.

Resetting mechanisms

Volume diffusion moves species through a crystal lattice and can produce geometry-dependent profiles. Grain-boundary diffusion or fracture transport shortens pathways. Dissolution-reprecipitation can replace a domain rapidly relative to lattice diffusion and preserve sharp reaction fronts or new porous material. Recrystallisation creates new grains and boundaries. Deformation introduces defects, subgrains and pathways. Fluid interaction changes chemical potential and removes or supplies components.

These mechanisms predict different evidence. A smooth core-rim gradient that scales with grain size may support diffusion. Sharp compositional and age boundaries associated with porosity or new mineral chemistry may support replacement. Patchy domains along cracks suggest fluid access. No texture is decisive alone, but mechanism-specific predictions make the alternatives testable.

Resetting need not be complete. Partial loss can move an analysis along a discordance relation or broaden an age spectrum. A later event may reset one mineral but not another, or rims but not cores. This difference is useful thermal and fluid-history evidence, not a reason to discard the younger values automatically.

Sampling and model design

Image grain size, shape, zoning, fractures, inclusions, damage and alteration before dating. Preserve the analysed coordinates and interaction volume. Measure compositional variables known to influence diffusion where possible. Avoid comparing closure behaviour from grains of different size or damage without including those factors.

Build thermal histories from independent geology: intrusion relations, metamorphic assemblages, structural sequence, sedimentary burial, thermochronology and thermal modelling. Cooling rate should be a range or path, not one unexamined number. Heating duration matters: a short high-temperature pulse may affect small or damaged domains while leaving large cores.

Use forward models first. Predict age or concentration profiles for candidate thermal and reaction histories, then compare with data. Inverse fitting can find several histories that reproduce the same dates. Report resolution and non-uniqueness rather than a single visually smooth path.

Worked synthetic example

Assume a constant illustrative diffusivity D=1.0\times10^{-22}\ \mathrm{m^2/s} for one stage lasting 1.00 Ma. Converting time gives t=3.156\times10^{13} s. The characteristic length is

$\ell=\sqrt{Dt} =\sqrt{(1.0\times10^{-22})(3.156\times10^{13})} =5.62\times10^{-5}\ \mathrm{m}=56\ \mu\mathrm{m}.$

A spherical domain with 50 \mum radius could be extensively affected under the assumed boundary condition; a 250 \mum radius grain might preserve a substantial core. This is a scaling result, not a complete diffusion solution. Real D changes with temperature, and surface concentration, geometry and daughter production must be included.

Now consider two grains that yield 420 and 690 Ma, while an imaged core yields 1,100 Ma. A simple weighted mean would erase the pattern. A better model tests whether grain size, fractures and rim thickness correlate with apparent age. If the smallest or most fractured domains are youngest and boundaries show replacement, fluid-assisted reaction may be more plausible than uniform volume diffusion. The dates then constrain a multi-stage history rather than one event.

Interpretation workflow and uncertainty

  1. Define which isotope species and mineral domain can exchange.
  2. Map grain geometry, damage, zoning, alteration and fast pathways.
  3. Select diffusion or reaction mechanisms from testable textural predictions.
  4. Record parameter sources, composition and calibration range.
  5. Construct bounded thermal and fluid histories from independent evidence.
  6. Forward-model profiles and domain ages before inversion.
  7. Compare grains of different size and preservation.
  8. Separate analytical uncertainty from kinetic parameters and history non-uniqueness.
  9. Report the event boundary actually constrained: growth, cooling, exchange or replacement.

Failure modes

Common failures include assigning a fixed closure temperature to a mineral name, using visible grain size instead of diffusion domain, ignoring anisotropy and damage, treating all resetting as volume diffusion, and inferring a unique cooling path from two dates. Another is deleting altered domains even when they record the event being investigated.

Kinetic parameter uncertainty can be orders of magnitude, and thermal histories are often correlated. Geological mechanism uncertainty may dominate. Present families of admissible histories and specify which new observation would reduce them.

Practice and review

  1. Calculate diffusion lengths for three values of D and two durations, then compare them with 25, 100 and 500 \mum domains.
  2. Sketch profiles predicted by volume diffusion, sharp replacement and fracture-controlled exchange.
  3. Explain how faster cooling, larger grain size and higher activation energy affect closure qualitatively.
  4. Design a sampling strategy that tests whether age depends on grain size or alteration.
  5. Given three mineral systems with different retentivity, write two thermal histories that could produce the same age order.

For the completion exercise, create a synthetic zoned-grain dataset with ages, radii, fractures and alteration scores. Test a diffusion-length model and a replacement model, preserve both where unresolved, and identify the next discriminating observation.

Sources and further reading